{ "architectures": [ "DomynEdgeForCausalLM" ], "attention_dropout": 0.0, "auto_map": { "AutoConfig": "configuration_domynedge.DomynEdgeConfig", "AutoModel": "modeling_domynedge.DomynEdgeModel", "AutoModelForCausalLM": "modeling_domynedge.DomynEdgeForCausalLM" }, "dtype": "bfloat16", "hidden_act": "silu", "hidden_size": 2560, "intermediate_size": 9728, "max_position_embeddings": 262144, "model_type": "domynedge", "num_attention_heads": 32, "num_hidden_layers": 36, "num_key_value_heads": 8, "rms_norm_eps": 1e-06, "position_layers": [ "nope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "rope", "rope", "rope", "nope", "nope", "nope" ], "window_sizes": [ [ -1, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ 1024, 0 ], [ -1, 0 ], [ -1, 0 ], [ -1, 0 ] ], "head_sizes": [ 256, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 64, 64, 64, 256, 256, 256 ], "tie_word_embeddings": true, "rope_theta": 10000.0, "transformers_version": "5.8.1", "use_cache": true, "vocab_size": 115272 }